Image Classification on ImageNet-1K 1.0 (val) (Top-1 Acc and Training Time)
81.8Top-1 AccuracyBaseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BaselineModel=DeiT-Base, Hyper-parameter consistency=true, Architecture consistency=true, Strategy consistency=true2024.03 | 81.8 | — | 292.8 | |
| ToE r1=0.5Model=DeiT-Base, Hyper-parameter consistency=true, Architecture consistency=true, Strategy consistency=true2024.03 | 81.6 | — | 231.2 | |
| NetworkExpansion 6→12Model=DeiT-Base, Hyper-parameter consistency=true, Architecture consistency=false, Strategy consistency=false2024.03 | 81 | 81.5 | 226.8 | |
| StackBERTModel=DeiT-Base, Hyper-parameter consistency=false, Architecture consistency=true, Strategy consistency=false2024.03 | 80.8 | — | 231.6 | |
| ToE r1=0.5Model=DeiT-Tiny, Hyper-parameter consistency=true, Architecture consistency=true, Strategy consistency=true2024.03 | 72.6 | — | 44.2 | |
| BaselineModel=DeiT-Tiny, Hyper-parameter consistency=true, Architecture consistency=true, Strategy consistency=true2024.03 | 72.2 | — | 54.6 | |
| ToMeModel=DeiT-Tiny, Hyper-parameter consistency=false, Architecture consistency=true, Strategy consistency=false2024.03 | 71.7 | — | 53.3 | |
| NetworkExpansion 6→12Model=DeiT-Tiny, Hyper-parameter consistency=true, Architecture consistency=false, Strategy consistency=false2024.03 | 70.3 | 70.1 | 43.2 | |
| S2ViTEModel=DeiT-Tiny, Training epochs=600, Hyper-parameter consistency=false, Architecture consistency=false, Strategy consistency=false2024.03 | 70.1 | — | — |